A Real-Time Vehicle Target Detection and Obstacle Avoidance Method and System Based on Edge Computing

Abstract

This invention relates to a real-time vehicle target detection and obstacle avoidance method and system based on edge computing. The method includes: an edge server acquiring vehicle information data transmitted from onboard devices in the vehicle in real time; the edge server processing the vehicle information data using a state machine model and an obstacle avoidance algorithm to obtain obstacle avoidance decision instructions; the edge server sending the obstacle avoidance decision instructions to the vehicle, and the vehicle executing the instructions; each time the vehicle completes an instruction, it sends a data packet reporting the execution status to the edge server to inform the server whether the current instruction was successfully executed; and the vehicle information data and obstacle avoidance decision instructions are updated in real time until the vehicle safely and correctly reaches its destination. Compared with existing technologies, this invention has advantages such as strong robustness and strong real-time performance.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to a real-time vehicle target detection and obstacle avoidance method and system based on edge computing. Background Technology

[0002] Current mainstream obstacle avoidance systems all use vehicles as computing platforms, relying on the vehicle's own computing power to collect and analyze data, and make decisions to achieve obstacle avoidance. This vehicle-centric computing solution places high demands on the vehicle's computing capabilities, increasing vehicle costs. Furthermore, as an onboard computer, the obstacle avoidance system presents challenges in upgrading and updating, making full-range version updates impossible. Running an outdated obstacle avoidance system may pose certain safety hazards to the vehicle.

[0003] In addition, there are technical solutions that rely on cloud computing servers as the computing platform. These solutions wirelessly transmit data and commands between vehicles and remote data center cloud computing servers. The vehicles primarily collect data, while the servers perform calculations and make decisions based on the transmitted data. However, this solution, which focuses on a data center cloud computing server platform, faces challenges due to the typically long distances between vehicles and servers, resulting in relatively high latency and less real-time performance. Furthermore, cloud computing servers are relatively large, leading to high management costs, and the centralized storage of user information on cloud servers makes them vulnerable to data breaches if attacked. Summary of the Invention

[0004] The purpose of this invention is to provide a real-time vehicle target detection and obstacle avoidance method and system based on edge computing. The method acquires data on the vehicle through image and ultrasonic ranging, and sends it to an edge server near the vehicle for calculation and decision-making, thereby achieving real-time obstacle avoidance for the vehicle with low requirements for vehicle computing power.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A real-time vehicle target detection and obstacle avoidance method based on edge computing includes the following steps:

[0007] The edge server acquires vehicle information data transmitted by the vehicle's onboard equipment in real time.

[0008] The edge server uses a state machine model to process vehicle information data based on an obstacle avoidance algorithm to obtain obstacle avoidance decision instructions. The state machine model allows the edge server to store part of the data after each processing for use in the next calculation, thereby obtaining a series of obstacle avoidance decision instructions.

[0009] The edge server sends obstacle avoidance decision instructions to the vehicle, and the vehicle executes the instructions;

[0010] Each time a vehicle completes an instruction, it sends a data packet to the edge server to report the execution status, informing the server whether the instruction was executed successfully or not.

[0011] The system updates vehicle information data and obstacle avoidance decision instructions in real time until the vehicle safely and correctly reaches its destination.

[0012] The vehicle information data includes data packet ID, total length, offset, data length, obstacle distance, and image data.

[0013] The obstacle avoidance algorithm includes the following steps:

[0014] Determine if the distance to the obstacle is less than the pre-configured obstacle avoidance distance. If not, set a forward command. If yes, determine if it is less than the pre-configured safe distance. If it is less than the safe distance, issue a reverse command until the vehicle reverses to the safe distance. Otherwise, execute the obstacle avoidance sub-process.

[0015] The obstacle avoidance sub-process includes the following steps:

[0016] Based on the object distance and image data, and using a deep learning network, the position of the obstacle in the camera coordinate system is determined, and the size of the obstacle is calculated.

[0017] Traverse all obstacles on the image and calculate the feasible area for the vehicle;

[0018] Calculate the bias and steering angle based on the feasible area, and control the vehicle to move in accordance with the bias and steering angle;

[0019] Determine if the current distance is less than the safe distance. If so, issue a reverse command until the vehicle reverses to the safe distance. Otherwise, determine if there is a next obstacle. If there is, move to the next obstacle and recursively execute the obstacle avoidance algorithm. If there is no obstacle, issue a command to return the vehicle to center and end the obstacle avoidance sub-process.

[0020] The selection of edge servers is determined based on an inter-server election algorithm. If there is only one edge server within the vehicle's pre-configured distance range, then this edge server is selected as the best server to provide services to the vehicle. If there are two or more servers, then one of them is selected as the best server to provide services to the vehicle.

[0021] The inter-server election algorithm includes the following steps:

[0022] Start all servers and run the program;

[0023] Determine the primary and secondary servers. If there is only one edge server, then that server is the primary server. The primary server acts as a TCP server, maintaining the connection between the vehicle and the secondary server, and assigning IDs to the vehicle and the secondary server. The secondary server receives the vehicle's request information, processes it, generates key parameters, and sends them to the primary server.

[0024] The main server receives the vehicle's request information and generates its own parameters;

[0025] When only the master server exists, it is determined whether the master server has received the request information for the first time. If so, the master server is the best server and sends its own IP response information to the vehicle. Otherwise, the master server continues to wait to receive the vehicle's request information.

[0026] When a secondary server exists, the primary server goes into sleep mode and waits for the secondary server to send key parameters. The key parameters are placed in the corresponding positions according to the vehicle and secondary server IDs. All server parameters are processed to select the best server. It is then determined whether the best server is currently providing services to the vehicle. If so, the primary server continues to accept the vehicle's waiting request information. If not, the IP response information of the best server is sent to the vehicle, and the primary server continues to wait to receive the vehicle's request information.

[0027] A real-time vehicle target detection and obstacle avoidance system based on edge computing includes an edge server, which comprises a data interaction module and an obstacle avoidance command determination module.

[0028] The data interaction module is used to perform the following steps: real-time acquisition of vehicle information data transmitted by the on-board equipment in the vehicle; sending obstacle avoidance decision commands to the vehicle; and receiving data packets reporting the execution status sent by the vehicle after executing the obstacle avoidance decision commands.

[0029] The obstacle avoidance command determination module adopts a state machine model and processes vehicle information data based on the obstacle avoidance algorithm to obtain obstacle avoidance decision commands. The state machine model enables the edge server to store part of the data after each processing for use in the next calculation, thereby obtaining a series of obstacle avoidance decision commands. The obstacle avoidance decision commands are updated in real time according to the real-time vehicle information data.

[0030] The vehicle information data includes data packet ID, total length, offset, data length, obstacle distance, and image data.

[0031] The obstacle avoidance algorithm includes the following steps:

[0032] Determine if the distance to the obstacle is less than the pre-configured obstacle avoidance distance. If not, set a forward command. If yes, determine if it is less than the pre-configured safe distance. If it is less than the safe distance, issue a reverse command until the vehicle reverses to the safe distance. Otherwise, execute the obstacle avoidance sub-process.

[0033] The obstacle avoidance sub-process includes the following steps:

[0034] Based on the object distance and image data, and using a deep learning network, the position of the obstacle in the camera coordinate system is determined, and the size of the obstacle is calculated.

[0035] Traverse all obstacles on the image and calculate the feasible area for the vehicle;

[0036] Calculate the bias and steering angle based on the feasible area, and control the vehicle to move in accordance with the bias and steering angle;

[0037] Determine if the current distance is less than the safe distance. If so, issue a reverse command until the vehicle reverses to the safe distance. Otherwise, determine if there is a next obstacle. If there is, move to the next obstacle and recursively execute the obstacle avoidance algorithm. If there is no obstacle, issue a command to return the vehicle to center and end the obstacle avoidance sub-process.

[0038] The selection of edge servers is determined based on an inter-server election algorithm. If there is only one edge server within the vehicle's pre-configured distance range, then this edge server is selected as the best server to provide services to the vehicle. If there are two or more servers, then one of them is selected as the best server to provide services to the vehicle.

[0039] The inter-server election algorithm includes the following steps:

[0040] Start all servers and run the program;

[0041] Determine the primary and secondary servers. If there is only one edge server, then that server is the primary server. The primary server acts as a TCP server, maintaining the connection between the vehicle and the secondary server, and assigning IDs to the vehicle and the secondary server. The secondary server receives the vehicle's request information, processes it, generates key parameters, and sends them to the primary server.

[0042] The main server receives the vehicle's request information and generates its own parameters;

[0043] When only the master server exists, it is determined whether the master server has received the request information for the first time. If so, the master server is the best server and sends its own IP response information to the vehicle. Otherwise, the master server continues to wait to receive the vehicle's request information.

[0044] When a secondary server exists, the primary server goes into sleep mode and waits for the secondary server to send key parameters. The key parameters are placed in the corresponding positions according to the vehicle and secondary server IDs. All server parameters are processed to select the best server. It is then determined whether the best server is currently providing services to the vehicle. If so, the primary server continues to accept the vehicle's waiting request information. If not, the IP response information of the best server is sent to the vehicle, and the primary server continues to wait to receive the vehicle's request information.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] (1) This invention enables vehicles and edge servers to play their respective roles and cooperate with each other by distributing tasks. Most of the main work is handed over to the edge server, which relieves the technical pressure on the vehicle. It has low requirements for the vehicle's computing power, strong robustness, and greatly improves efficiency.

[0047] (2) The present invention utilizes the working principle of the state machine model, where the next state depends on the previous state, making the control decision simpler, easier to understand, more reliable, and with lower latency.

[0048] (3) The present invention makes user data more secure to a certain extent by using device data distributed on edge servers.

[0049] (4) The present invention uses distributed edge computing to make the whole process less time-consuming and more real-time.

[0050] (5) The server election algorithm of the present invention can coordinate edge servers well. If the current edge server fails, the vehicle can be switched to another edge server without causing service interruption. Attached Figure Description

[0051] Figure 1 This is a flowchart of the method of the present invention;

[0052] Figure 2 This is a schematic diagram of the communication process between the vehicle and the edge server according to the present invention;

[0053] Figure 3 Here is the overall flowchart of the obstacle avoidance algorithm;

[0054] Figure 4 Obstacle avoidance flowchart;

[0055] Figure 5 This is a flowchart of the inter-server election process;

[0056] Figure 6 This is a schematic diagram of a state machine model. Detailed Implementation

[0057] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0058] A real-time vehicle target detection and obstacle avoidance method based on edge computing, such as Figure 1 As shown, it includes the following steps:

[0059] 1) The edge server acquires vehicle information data transmitted by the on-board equipment in the vehicle in real time.

[0060] The vehicle's onboard equipment senses the surrounding environment to acquire data such as images and distances. To ensure service quality, communication also occurs between the vehicle and the server. The vehicle information data includes data packet ID, total length, offset, data length, obstacle distance, and image data.

[0061] The communication process between the vehicle and the edge server is as follows: Figure 2 As shown.

[0062] 2) Edge servers adopt a state machine model, such as... Figure 6 As shown, vehicle information data is processed based on an obstacle avoidance algorithm to obtain obstacle avoidance decision instructions. The state machine model enables the edge server to store part of the data after each processing for use in the next calculation, thereby obtaining a series of obstacle avoidance decision instructions.

[0063] The flowchart of the obstacle avoidance algorithm is as follows: Figure 3 As shown, it includes the following steps:

[0064] Determine if the distance to the obstacle is less than the pre-configured obstacle avoidance distance. If not, set a forward command. If yes, determine if it is less than the pre-configured safe distance. If it is less than the safe distance, issue a reverse command until the vehicle reverses to the safe distance. Otherwise, execute the obstacle avoidance sub-process.

[0065] The obstacle avoidance sub-process is as follows: Figure 4 As shown, it includes the following steps:

[0066] Based on the object distance and image data, and using a deep learning network, the position of the obstacle in the camera coordinate system is determined, and the size of the obstacle is calculated.

[0067] Traverse all obstacles on the image and calculate the feasible area for the vehicle;

[0068] Calculate the bias and steering angle based on the feasible area, and control the vehicle to move in accordance with the bias and steering angle;

[0069] Determine if the current distance is less than the safe distance. If so, issue a reverse command until the vehicle reverses to the safe distance. Otherwise, determine if there is a next obstacle. If there is, move to the next obstacle and recursively execute the obstacle avoidance algorithm. If there is no obstacle, issue a command to return the vehicle to center and end the obstacle avoidance sub-process.

[0070] 3) The edge server sends obstacle avoidance decision instructions to the vehicle, and the vehicle executes the instructions.

[0071] 4) After each instruction is executed, the vehicle sends a data packet to the edge server to report the execution status, informing the server whether the instruction was executed successfully.

[0072] 5) Update vehicle information data and obstacle avoidance decision instructions in real time until the vehicle safely and correctly reaches its destination.

[0073] The selection of edge servers is determined based on an inter-server election algorithm. If there is only one edge server within the vehicle's pre-configured distance range, then this edge server is selected as the best server to provide services to the vehicle. If there are two or more servers, then one of them is selected as the best server to provide services to the vehicle.

[0074] The inter-server election algorithm process is as follows: Figure 5 As shown, it includes the following steps:

[0075] Start all servers and run the program;

[0076] Determine the primary and secondary servers. If there is only one edge server, then that server is the primary server. The primary server acts as a TCP server, maintaining the connection between the vehicle and the secondary server, and assigning IDs to the vehicle and the secondary server. The secondary server receives the vehicle's request information, processes it, generates key parameters, and sends them to the primary server.

[0077] The main server receives the vehicle's request information and generates its own parameters;

[0078] When only the master server exists, it is determined whether the master server has received the request information for the first time. If so, the master server is the best server and sends its own IP response information to the vehicle. Otherwise, the master server continues to wait to receive the vehicle's request information.

[0079] When a secondary server exists, the primary server goes into sleep mode and waits for the secondary server to send key parameters. The key parameters are placed in the corresponding positions according to the vehicle and secondary server IDs. All server parameters are processed to select the best server. It is then determined whether the best server is currently providing services to the vehicle. If so, the primary server continues to accept the vehicle's waiting request information. If not, the IP response information of the best server is sent to the vehicle, and the primary server continues to wait to receive the vehicle's request information.

[0080] Since the main server will accept vehicle requests and generate its own parameters regardless of the total number of edge servers, this setup ensures that even if a server goes offline due to a failure or other reasons, leaving only one server available, it can still identify a server to handle obstacle avoidance.

[0081] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A real-time vehicle target detection and obstacle avoidance method based on edge computing, characterized in that, Includes the following steps: The edge server acquires vehicle information data transmitted by the vehicle's onboard equipment in real time. The edge server uses a state machine model to process vehicle information data based on an obstacle avoidance algorithm to obtain obstacle avoidance decision instructions. The state machine model allows the edge server to store part of the data after each processing for use in the next calculation, thereby obtaining a series of obstacle avoidance decision instructions. The edge server sends obstacle avoidance decision instructions to the vehicle, and the vehicle executes the instructions; Each time a vehicle completes an instruction, it sends a data packet to the edge server to report the execution status, informing the server whether the instruction was executed successfully or not. The system updates vehicle information data and obstacle avoidance decision instructions in real time until the vehicle safely and correctly reaches its destination. The selection of edge servers is determined based on an inter-server election algorithm, which includes the following steps: Start all servers and run the program; Determine the primary and secondary servers. If there is only one edge server, then that server is the primary server. The primary server acts as a TCP server, maintaining the connection between the vehicle and the secondary server, and assigning IDs to the vehicle and the secondary server. The secondary server receives the vehicle's request information, processes it, generates key parameters, and sends them to the primary server. The main server receives the vehicle's request information and generates its own parameters; When only the master server exists, it is determined whether the master server has received the request information for the first time. If so, the master server is the best server and sends its own IP response information to the vehicle. Otherwise, the master server continues to wait to receive the vehicle's request information. When a secondary server exists, the primary server goes into sleep mode and waits for the secondary server to send key parameters. The key parameters are placed in the corresponding positions according to the vehicle and secondary server IDs. All server parameters are processed to select the best server. It is then determined whether the best server is currently providing services to the vehicle. If so, the primary server continues to accept the vehicle's waiting request information. If not, the IP response information of the best server is sent to the vehicle, and the primary server continues to wait to receive the vehicle's request information.

2. The real-time vehicle target detection and obstacle avoidance method based on edge computing according to claim 1, characterized in that, The vehicle information data includes data packet ID, total length, offset, data length, obstacle distance, and image data.

3. The real-time vehicle target detection and obstacle avoidance method based on edge computing according to claim 1, characterized in that, The obstacle avoidance algorithm includes the following steps: Determine if the distance to the obstacle is less than the pre-configured obstacle avoidance distance. If not, set a forward command. If yes, determine if it is less than the pre-configured safe distance. If it is less than the safe distance, issue a reverse command until the vehicle reverses to the safe distance. Otherwise, execute the obstacle avoidance sub-process. The obstacle avoidance sub-process includes the following steps: Based on the object distance and image data, and using a deep learning network, the position of the obstacle in the camera coordinate system is determined, and the size of the obstacle is calculated. Traverse all obstacles on the image and calculate the feasible area for the vehicle; Calculate the bias and steering angle based on the feasible area, and control the vehicle to move in accordance with the bias and steering angle; Determine if the current distance is less than the safe distance. If so, issue a reverse command until the vehicle reverses to the safe distance. Otherwise, determine if there is a next obstacle. If there is, move to the next obstacle and recursively execute the obstacle avoidance algorithm. If there is no obstacle, issue a command to return the vehicle to center and end the obstacle avoidance sub-process.

4. The real-time vehicle target detection and obstacle avoidance method based on edge computing according to claim 1, characterized in that, If there is only one edge server within the vehicle's pre-configured distance range, then this edge server is selected as the best server to provide services to the vehicle; if there are two or more servers, then one of them is selected as the best server to provide services to the vehicle.

5. A real-time vehicle target detection and obstacle avoidance system based on edge computing, characterized in that, It includes an edge server, which comprises a data interaction module and an obstacle avoidance command determination module, wherein... The data interaction module is used to perform the following steps: real-time acquisition of vehicle information data transmitted by the on-board equipment in the vehicle; sending obstacle avoidance decision commands to the vehicle; and receiving data packets reporting the execution status sent by the vehicle after executing the obstacle avoidance decision commands. The obstacle avoidance command determination module adopts a state machine model and processes vehicle information data based on the obstacle avoidance algorithm to obtain obstacle avoidance decision commands. The state machine model enables the edge server to store part of the data after each processing for use in the next calculation, thereby obtaining a series of obstacle avoidance decision commands. The obstacle avoidance decision commands are updated in real time according to the real-time vehicle information data. The selection of edge servers is determined based on an inter-server election algorithm, which includes the following steps: Start all servers and run the program; Determine the primary and secondary servers. If there is only one edge server, then that server is the primary server. The primary server acts as a TCP server, maintaining the connection between the vehicle and the secondary server, and assigning IDs to the vehicle and the secondary server. The secondary server receives the vehicle's request information, processes it, generates key parameters, and sends them to the primary server. The main server receives the vehicle's request information and generates its own parameters; When only the master server exists, it is determined whether the master server has received the request information for the first time. If so, the master server is the best server and sends its own IP response information to the vehicle. Otherwise, the master server continues to wait to receive the vehicle's request information. When a secondary server exists, the primary server goes into sleep mode and waits for the secondary server to send key parameters. The key parameters are placed in the corresponding positions according to the vehicle and secondary server IDs. All server parameters are processed to select the best server. It is then determined whether the best server is currently providing services to the vehicle. If so, the primary server continues to accept the vehicle's waiting request information. If not, the IP response information of the best server is sent to the vehicle, and the primary server continues to wait to receive the vehicle's request information.

6. The real-time vehicle target detection and obstacle avoidance system based on edge computing according to claim 5, characterized in that, The vehicle information data includes data packet ID, total length, offset, data length, obstacle distance, and image data.

7. A real-time vehicle target detection and obstacle avoidance system based on edge computing according to claim 5, characterized in that, The obstacle avoidance algorithm includes the following steps: Determine if the distance to the obstacle is less than the pre-configured obstacle avoidance distance. If not, set a forward command. If yes, determine if it is less than the pre-configured safe distance. If it is less than the safe distance, issue a reverse command until the vehicle reverses to the safe distance. Otherwise, execute the obstacle avoidance sub-process. The obstacle avoidance sub-process includes the following steps: Based on the object distance and image data, and using a deep learning network, the position of the obstacle in the camera coordinate system is determined, and the size of the obstacle is calculated. Traverse all obstacles on the image and calculate the feasible area for the vehicle; Calculate the bias and steering angle based on the feasible area, and control the vehicle to move in accordance with the bias and steering angle; Determine if the current distance is less than the safe distance. If so, issue a reverse command until the vehicle reverses to the safe distance. Otherwise, determine if there is a next obstacle. If there is, move to the next obstacle and recursively execute the obstacle avoidance algorithm. If there is no obstacle, issue a command to return the vehicle to center and end the obstacle avoidance sub-process.

8. A real-time vehicle target detection and obstacle avoidance system based on edge computing according to claim 5, characterized in that, If there is only one edge server within the vehicle's pre-configured distance range, then this edge server is selected as the best server to provide services to the vehicle; if there are two or more servers, then one of them is selected as the best server to provide services to the vehicle.

Images